EXPLORE 300: Exploring Adobe Journey Optimizer System Datasets with Data Distiller

Unleashing Insights from Adobe Journey Optimizer Datasets with Data Distiller

Prerequisites

You need a basic understanding of how to write nested queries and working with nested data.

You should get familiar with navigating around with web data:

Overview

To generate a holistic view of how different datasets contribute to serving an experience via Adobe Journey Optimizer (AJO), I will now walk through each dataset in the correct order of importance and process flow. These datasets include:

1. AJO Entity Record Schema Dataset with the dataset name: ajo_event_dataset

2. Journey Step Events Dataset with the dataset name:journey_step_events

3. AJO Message Feedback Events Dataset with the dataset namee: ajo_message_feedback_event_dataset

4. BCC Feedback Events Dataset with the dataset name: ajo_bcc_feedback_event_dataset

5. AJO Email & Push Tracking Datasets with the dataset names: ajo_email_tracking_experience_event_dataset, ajo_push_tracking_experience_event_dataset

6. Offer Decisioning Events Dataset with the dataset name: ode_decisionevents_{key specific to your environment}.

The chapter contains additional 4 Offer Decisioning datasets that give you deeper information about the offers and the decisioning logic.

How the Datasets Work Together in Adobe Journey Optimizer (AJO)

In Adobe Journey Optimizer, each dataset serves a specific role in orchestrating, delivering, and optimizing customer experiences. When combined, these datasets provide a comprehensive understanding of how customer journeys are executed, how messages are delivered and engaged with, and how offers are decided and optimized.

Here’s how each dataset is related, presented in the correct order of importance and process flow:

1. AJO Entity Record Schema Dataset: The Core Foundation

• Purpose: The AJO Entity Record Schema Dataset is the central dataset that logs and tracks the metadata for all journeys. It captures crucial information about the campaign, messages, journey actions, and message triggers. It forms the basis for connecting all other datasets in the system.

• Role in the Process:

  • Journey Orchestration: This dataset logs the entire structure of the journey, including message triggers, campaign actions, journey steps, and decisions.

  • It includes identifiers like Message IDs, Campaign IDs, and Journey Action IDs, which link to the Message Feedback, Tracking, and ODE datasets.

  • Without this dataset, none of the other datasets would have the necessary context to operate. It establishes the backbone of the journey and ensures that all steps are executed as per the designed journey.

2. Journey Step Events Dataset: Tracking Journey Progression

• Purpose: The Journey Step Events Dataset provides detailed insights into each step within the journey. It logs step-level events, including step completions, errors, timeouts, and transitions. This dataset ensures visibility into how users progress through the journey and helps diagnose any issues.

• Role in the Process:

  • Step-Level Monitoring: This dataset records each step a user takes, whether that step is completed successfully, if there are errors, or if a journey action times out.

  • Action Execution: It tracks the execution of actions (such as sending an email or showing an offer) and logs the results of those actions.

  • Error Handling: Any errors encountered during journey execution are logged, helping you resolve issues at specific steps.

  • Relation to Other Datasets: The Journey Step Events Dataset links to the AJO Entity Record Schema and the ODE Dataset to ensure that each decision or action triggered within the journey is properly tracked and logged.

3. AJO Message Feedback Events Dataset: Delivery Tracking

• Purpose: The Message Feedback Events Dataset focuses on delivery feedback for emails, SMS, and push notifications. It logs the delivery status, including whether the message was delivered, bounced, or required retries.

• Role in the Process:

  • Delivery Status Monitoring: After a message is triggered by a journey step (as logged in the Journey Step Events Dataset), the Message Feedback Events Dataset tracks whether the message was delivered successfully or encountered a failure.

  • Bounce & Failure Tracking: It logs details such as bounce reasons, invalid emails, or retries, providing insight into delivery issues and helping you troubleshoot any problems with sending.

  • Relation to Other Datasets: The Message Feedback Dataset ties back to the AJO Entity Record Schema via the Message ID, ensuring that the status of every message triggered by the journey is accounted for.

4. BCC Feedback Events Dataset: Tracking Secondary Recipients

• Purpose: The BCC Feedback Events Dataset tracks the delivery status of emails sent to BCC (Blind Carbon Copy) or CC recipients. This dataset is important for ensuring compliance and tracking delivery to these secondary recipients.

• Role in the Process:

  • Secondary Delivery Monitoring: For messages sent to BCC or CC recipients (often for compliance or archiving purposes), this dataset logs the delivery status and captures whether these secondary emails were successfully delivered or excluded.

  • Exclusion Handling: It tracks exclusions due to compliance rules or typology filters and provides insight into why certain emails were excluded.

  • Relation to Other Datasets: Like the Message Feedback Events Dataset, it ties back to the AJO Entity Record Schema to track secondary recipients, ensuring full coverage of all recipients in the system.

5. AJO Email & Push Tracking Datasets: User Engagement

• Purpose: The Tracking Datasets for email and push notifications log user engagement with delivered messages, including metrics such as opens, clicks, and unsubscribes. This dataset helps measure the effectiveness of the messages after they are successfully delivered.

• Role in the Process:

  • Engagement Monitoring: Once a message is delivered (tracked via the Message Feedback Dataset), the Tracking Datasets log how users interact with that message—whether they open it, click on a link, or unsubscribe.

  • Performance Reporting: These datasets provide insights into how well messages perform in terms of user engagement and can be used to optimize future campaigns based on click-through rates and engagement metrics.

  • Relation to Other Datasets: The Tracking Datasets link back to the Message Feedback Dataset and the AJO Entity Record Schema via the Message ID, ensuring that you have a full picture of the message’s journey from delivery to engagement.

6. Offer Decisioning Events Dataset: Optimizing Decision-Making

• Purpose: The Offer Decisioning Events Dataset tracks decision points within the journey where offers are presented to users. It logs which offers were shown, how users interacted with them (e.g., clicks or conversions). It logs decisions made during the journey based on rules, algorithms, or fallback options.

• Role in the Process:

  • Decision Tracking: When a decision point in the journey is reached, this dataset logs which offer was selected and whether the user engaged with it.

  • Optimization of Decision Strategies: By tracking offer performance, you can analyze which offers perform best, optimize decision strategies, and refine the algorithms used to present offers.

  • Relation to Other Datasets: The Offer Decisioning Events Dataset connects with the Journey Step Events Dataset to log when a decision point was triggered and which offer was selected. It is also tied to the AJO Entity Record Schema to ensure that decisions made within the journey are fully tracked.

Bringing It All Together: End-to-End Experience Monitoring in AJO

1. Journey Setup and Execution (AJO Entity Record Schema Dataset & Journey Step Events Dataset):

   • The AJO Entity Record Schema Dataset forms the foundation for the entire journey, logging messages, actions, and decisions taken within the journey.

   • The Journey Step Events Dataset tracks each step in the journey, ensuring that actions like sending a message or making a decision are logged and monitored for performance and errors.

2. Message Delivery (Message Feedback Events Dataset & BCC Feedback Events Dataset):

   • After a message is triggered in the journey, the Message Feedback Events Dataset tracks whether the message was successfully delivered or bounced.

   • The BCC Feedback Events Dataset tracks the status of BCC and CC recipients, ensuring that secondary recipients are handled properly and that compliance requirements are met.

3. User Engagement (AJO Email & Push Tracking Datasets):

   • Once a message is delivered, the Tracking Datasets capture user engagement, including opens, clicks, and unsubscribes. This data provides insights into the effectiveness of messages in driving user behavior.

4. Offer Decisioning and Optimization (Offer Decisioning Events Dataset):

   • Throughout the journey, decisions are made regarding which offers to present to users. The Offer Decisioning Events Dataset logs these decisions, tracks offer engagement, and helps you optimize your decision-making strategies.

How to Use the Datasets Together:

• Monitor Journey Progress: Use the AJO Entity Record Schema Dataset and Journey Step Events Dataset to monitor the overall progress and structure of the customer journey. These datasets help you track which steps were taken and whether any issues occurred.

• Ensure Message Delivery: Leverage the Message Feedback Events Dataset and BCC Feedback Events Dataset to track whether messages triggered by the journey were successfully delivered, and identify any bounces or failures.

• Analyze Engagement: After messages are delivered, use the Tracking Datasets to analyze user engagement and optimize future campaigns based on how users interacted with the message.

• Optimize Offer Decisions: Use the Offer Decisioning Events Dataset to analyze which offers were presented to users

Schema Dictionary for AJO System Datasets

You can find the exhaustive list here.

AJO Entity Record Schema Dataset

First, execute the following query in the Data Distiller Query Pro Mode Editor:

SELECT 
   table_name, 
   column_name, 
   data_type 
FROM 
   information_schema.columns
where table_name = 'ajo_entity_dataset'

The result will be:

Now execute:

SELECT to_json(_experience) FROM ajo_entity_dataset LIMIT 500;

The result will be:

The AJO Entity Record Schema is designed to store metadata related to messages sent to end-users within Adobe Journey Optimizer (AJO). It captures essential data related to campaigns, journeys, channels (email, SMS, push notifications), and experiments. This schema is integral for tracking and analyzing campaign performance, engagement, conversions, and message delivery across various channels. Think of this dataset acting as timestamped lookup dataset for all the otyher datasets that contain tracking and feedback information on the messages that was sent out. The lookup data is timestamp as the metadata can change as a function of time with users making changes to the various configurations.

You cannot use event-specific identifiers like _id and timestamp, as they are tied to the logging of individual events. Therefore, your best option is to link the message IDs together. The messageID attribute in every record in this dataset is absolutely critical because it helps to stitch various datasets such as Message Feedback Dataset and Experience Event Tracking Datasets to get details of a message delivery from sending to tracking at a profile level. An entry for a message is created only after journey or campaign is published. You may see the entry/update 30 minutes after the publication of the campaign/journey.

Since the AJO Entity Record Schema is the central lookup for all the other datasets, this field in the dataset ajo_entity_dataset

_experience.customerJourneyManagement.entities.channelDetails.messageID

Here are the key fields that you need to be aware of:

Journey Step Event Dataset

You should be able to execute the following code:

SELECT * FROM journey_step_events LIMIT 500;

The Journey Step Event Dataset in Adobe Journey Optimizer captures and logs all journey step experience events as part of Journey Orchestration. These events are essential for reporting and analytics in systems like Customer Journey Analytics. The dataset helps track each step within a journey and its performance, providing insights into how users progress through their customer journey, how actions are executed, and what the results of those actions are. This dataset is especially useful for understanding step-level events within journeys, such as errors, transitions, and completions.

Key Use Cases for the Journey Step Event Dataset:

1. Journey Reporting and Analysis: Provides visibility into the execution and performance of individual steps within journeys, such as transitions between steps, completion rates, and timeouts.

2. Error Tracking and Resolution: Logs errors and failure codes associated with journey steps, helping diagnose and resolve issues that affect customer experience.

3. Journey Optimization: Tracks how users move through the journey, allowing marketers to optimize step transitions, messaging timing, and action results for better engagement.

4. Profile Segmentation and Interaction: Captures profile identifiers and segment qualifications, which are essential for targeting and personalizing the user journey.

Here are the key fields and the unique ones are in orange:

The Segment ID field is found in the Journey Step Events Dataset.

• The field path for Segment ID is _experience.journeyOrchestration.stepEvents.segmentExportJob.exportSegmentID.

• This field captures the segment identifier when a segment export job is triggered during the journey orchestration process.

This is critical for understanding which segment was used during a particular step of the journey, especially in journeys that are triggered by audience segments. This information allows you to link specific segment behaviors with journey events, providing detailed insights into how segment membership affects journey progression and outcomes.

AJO Message Feedback Datasets

Focus: Primarily focused on feedback from ISPs or service providers after an attempt to deliver a message (email, SMS, or push).

First, go ahead and execute this:

SELECT * FROM ajo_message_feedback_event_dataset LIMIT 500;

The AJO Message Feedback Event Dataset is a dataset designed to log and track the delivery of messages within Adobe Journey Optimizer (AJO). It provides comprehensive feedback on message delivery attempts across multiple channels such as email, push notifications, and SMS:

• Logs detailed delivery information, including bounces, retry attempts, failure reasons, and status (delivered, failed, etc.).

• Provides diagnostic feedback on why a message succeeded or failed, helping improve deliverability.

• Focuses on the message journey from the system to the recipient’s inbox or device.

• Captures feedback regarding message delivery failure (e.g., async bounce, sync bounce, invalid email address).

Key Use Cases:

• Delivery Status Reporting: Detailed insights into delivery success and failure.

• Bounce and Retry Analysis: Helps diagnose why messages failed and how many retry attempts were made.

• Compliance and Monitoring: Tracks outbound IP addresses, bounce types, and reasons for failures.

Here are the fields that are most critical here. Note that the unique fields are in orange:

Why These Fields Are Important:

• Delivery Status & Failure Reason: These fields are crucial for understanding message delivery success and failure, as well as diagnosing the reasons behind message bounces and undelivered emails.

• Retry Count: Helps analyze retry behavior and can reveal patterns in which retry attempts are successful and which are not.

• Offer & Proposition Data: Offer engagement tracking is essential to understanding how users interact with promotional content, enabling teams to optimize future campaigns based on conversion data.

• Journey Action ID: This links the message feedback back to the customer journey, providing insights into the effectiveness of different journey steps in triggering user engagement.

• Interaction Outcome: This field provides key insights into how recipients are interacting with the message, allowing for better tracking of conversion rates and user behavior following message delivery.

AJO Email BCC Feedback Event Dataset

First, execute the query:

SELECT * FROM ajo_bcc_feedback_event_dataset LIMIT 500;

The AJO Email BCC Feedback Event Dataset is specifically designed to track and log the delivery status of BCC (Blind Carbon Copy) emails. It is used primarily for reporting purposes to understand how BCC emails are handled, delivered, and processed, focusing on feedback such as exclusions, failures, and delivery outcomes.

Key Differences Between the BCC Feedback Event Dataset and the Message Feedback Event Dataset

1. BCC-specific Tracking: The BCC dataset is specifically focused on BCC and CC recipients, whereas the Message Feedback dataset logs information for all messages across email, SMS, and push channels. It includes fields for tracking the original recipient and the secondary recipient type (e.g., BCC, CC, Archival).

2. Exclusion Data: The BCC dataset contains fields like Exclusion Code and Exclusion Reason, which provide specific reasons for message exclusions, such as compliance or typology rules, which may not be as granular in the Message Feedback dataset.

3. Field Overlap: Both datasets share fields related to message delivery feedback, such as Delivery Status, Failure Category, Failure Reason, and Offer Information.

4. Use Case: The BCC Feedback Dataset is more narrowly focused on tracking BCC and CC email handling and is highly specialized for reporting purposes about those secondary recipients. The Message Feedback Dataset offers a broader scope, focusing on all message types across multiple channels (email, SMS, push), providing a wider range of delivery feedback, retries, and engagement.

Here are the key fields. Unique fields are marked in orange

Why These Fields Are Important:

• Delivery Status & Exclusion Data: These fields are key for understanding delivery performance and exclusion reasons, particularly when messages are filtered out by typology rules or compliance filters.

• Secondary Recipient Data: Unique to the BCC dataset, fields like Original Recipient Address and Secondary Recipient Type help track how secondary recipients (BCC, CC) are handled, which is critical for understanding email distribution and compliance.

• Offer & Proposition Data: These fields help measure the effectiveness of offers and promotions sent to BCC recipients, providing insights into engagement and offer performance.

AJO Email Tracking Experience Event Dataset

Focus: Concentrates on user interactions with delivered messages (email, SMS, push notifications)

Just type this query in the Data Distiller Query Pro Mode Editor:

SELECT * FROM ajo_email_tracking_experience_event_dataset LIMIT 500;

The results from above should be a great starting point for you to dig deeper into this dataset. The AJO Email Tracking Experience Event Dataset is designed to capture and log detailed interaction data related to email campaigns sent via the Adobe Journey Optimizer (AJO). This dataset tracks various user actions upon receiving emails, providing essential insights for performance reporting, segmentation, and optimization of email marketing campaigns:

1. Capturing User Interactions: The dataset records detailed information about how users interact with email campaigns, including:

   • Opens: Whether and how many times a recipient opened an email.

   • Clicks: Whether the recipient clicked on any links within the email.

   • Unsubscribes: Whether the user unsubscribed from future emails.

   • Bounces: Whether the email failed to be delivered (soft or hard bounce).

   • Deliveries: Logs whether the email was successfully delivered.

2. Email Performance Metrics: The dataset supports analysis of email performance with the following key metrics:

   • Open Rates: The percentage of recipients who opened the email, useful for assessing the effectiveness of subject lines.

   • Click-Through Rates (CTR): The percentage of recipients who clicked on links within the email, indicating the relevance of the content or call-to-action (CTA).

   • Unsubscribe Rates: Tracks how many users opted out of future emails, helping to manage list hygiene and content relevance.

   • Bounce Rates: Identifies emails that were not delivered due to issues like invalid email addresses (hard bounces) or temporary issues (soft bounces).

3. Link and Offer Tracking: The dataset allows for detailed reporting on link and offer engagement, capturing:

   • Tracker URLs: Tracks the specific URLs that users clicked within the email.

   • Offer Interactions: Logs interactions with special offers or promotions included in the email, helping to measure the effectiveness of discounts, sales, or calls-to-action.

   • Landing Pages: Tracks if users landed on specific pages after clicking links, allowing for detailed conversion analysis.

4. Campaign and Journey Metadata: The dataset contains critical metadata regarding the email campaigns and journeys, including:

   • Campaign IDs: Unique identifiers for each campaign, enabling tracking of email performance across different campaigns.

   • Journey Action IDs: Tracks which specific journey actions triggered the email, useful for analyzing the effectiveness of different touchpoints.

   • Campaign Versioning: Enables the comparison of different versions of a campaign or journey to identify which versions are more effective.

5. Segmentation and Personalization: The dataset is enabled for profile integration, meaning it can be used for segmentation and personalized marketing:

   • Segment Creation: Build segments based on user behavior, such as frequent openers, non-clickers, or users who unsubscribed.

   • Personalization Insights: Analyze how different audience segments interact with emails, helping to tailor future campaigns for improved engagement.

6. Detailed Reporting for Compliance and Preference Management: The dataset helps track consent and compliance-related interactions, such as:

   • Email Preferences: Tracks user consent and opt-in preferences (e.g., GDPR compliance).

   • Unsubscribes: Provides information about users who opted out of future communications, ensuring adherence to privacy regulations.

7. A/B Testing and Optimization: The dataset supports A/B testing by tracking different email variants (e.g., subject lines, content, offers), allowing you to:

   • Test different variants: Measure how different content versions, send times, or calls-to-action perform to optimize future emails.

   • Send Time Optimization: Track whether send-time optimization strategies were applied, helping you to analyze the performance impact of different send times.

Reporting Use Cases:

• Performance Monitoring: Gain insight into how well email campaigns perform based on metrics such as opens, clicks, and conversions.

• Engagement Insights: Analyze how recipients interact with emails, including the most clicked links, offers, and CTAs.

• Conversion Tracking: Measure how well emails drive conversions, such as sales, sign-ups, or engagement with landing pages.

• A/B Testing: Compare the performance of different email versions to identify the most effective strategies.

• Deliverability and Bounce Analysis: Understand which emails failed to deliver and why, to optimize delivery rates and maintain list hygiene.

• Unsubscribe Management: Track and reduce unsubscribe rates by improving content relevance and targeting strategies.

Here are the fields that you will need. Fields marked in orange are unique to SMS notifications:

Key Interaction Types:

• Opens: Tracked through openCount and eventType for open events.

• Clicks: Measured using clickCount, trackerURL, and trackerURLLabel to see which links were clicked.

• Unsubscribes: The unsubscribed field records if a user opts out after receiving an email.

• Bounces: Captured through deliveryStatus and bounceType, detailing whether emails were delivered or bounced.

• Landing Page Engagement: landingPageID and landingPageName track which landing pages users visited after clicking links.

AJO Push Tracking Experience Event Dataset

Focus: Concentrates on user interactions with delivered messages (email, SMS, push notifications)

To explore this dataset, just type ad execute this in the Data Distiller Query Pro Mode Editor:

SELECT * FROM ajo_push_tracking_experience_event_dataset LIMIT 500;

The AJO Push Tracking Experience Event Dataset is designed to capture and log interaction events related to push notifications (including SMS) sent via the Adobe Journey Optimizer (AJO). This dataset stores detailed information about user interactions with push notifications, providing essential insights for reporting, segmentation, and performance analysis:

1. Capturing User Interactions: The dataset records various actions users take in response to push notifications, such as:

   • Receives: Whether the push notification was delivered to the user’s device.

   • Opens: Whether the user opened the app or interacted with the notification.

   • Clicks: Whether the user clicked any custom actions within the notification.

   • Dismisses: Whether the user dismissed the notification without engaging.

   • Launches: Whether the push notification successfully launched the app.

2. Push Notification Metadata: The dataset contains metadata about the push notifications, including:

   • Push Provider Information: Identifies which push provider (e.g., APNS for iOS, FCM for Android) was used to deliver the notification.

   • Push Provider Message ID: Unique identifier assigned to the notification by the provider.

   • Custom Actions: Logs any custom actions (e.g., buttons) included in the push notification and records user interactions with them.

3. Tracking User Engagement: Information in the dataset supports the measurement of key performance indicators such as:

   • Open rates: The percentage of users who open or interact with push notifications.

   • Engagement rates: Based on custom action clicks or other interactions within the notification.

   • Conversion: If push notifications prompt specific user actions, such as purchases or sign-ups within the app.

4. Segmentation and Profiling: The dataset is enabled for profile integration, meaning it can be used to build audience segments based on user interaction data. For example:

   • Segment users who frequently open push notifications.

   • Target users who never engage with notifications.

   • Measure user engagement with specific campaigns to refine marketing strategies.

5. Supporting Campaign Analysis: It includes detailed information about the campaigns and journeys that trigger push notifications, such as:

   • Campaign IDs: Track push notification performance by campaign.

   • Journey Action IDs: Helps identify which journey action led to the notification being sent.

   • Journey Versioning: Enables performance comparison between different versions of journeys or campaigns.

6. Geolocation and Contextual Data: For use cases involving location-based push notifications, the dataset can capture contextual data such as:

   • Geo-location data: Logs when notifications are triggered by location-based events (e.g., entering a specific geographical area).

   • Points of Interest (POIs): Logs interaction with POIs when they are used to trigger notifications.

Reporting Use Cases:

• Performance Monitoring: Understand how different push notifications perform across various campaigns and journeys.

• Engagement Insights: Track how users interact with notifications, including opens, custom action clicks, and app launches.

• Conversion Tracking: Measure how effective push notifications are at driving conversions, such as app launches or purchases.

• A/B Testing: Compare different versions of push notifications to see which variants (message types, delivery times, custom actions) perform better.

Push

Here are the fields that you will need. Fields marked in orange are unique to push notifications:

SMS

Here are the fields that you will need. Fields marked in orange are unique to SMS notifications:

Offer Decisions Events Dataset

First, you need to execute the following by locating the dataset that has ode_decisionevents in its name:

SELECT * FROM ode_decisionevents_{key specific to your environment};

A proposition offer is a specific type of personalized offer or recommendation presented to a customer during their journey in Adobe Journey Optimizer (AJO) or Adobe Experience Platform (AEP). It can be anything from a product recommendation, discount, or special promotion that is generated based on a user’s behavior, preferences, or profile data. The proposition offer is intended to drive engagement, conversion, or retention by aligning with the user’s interests and needs. A decision is the process by which the system determines what action or offer to present to a user based on a set of rules, algorithms, or predefined criteria. It is a critical part of personalized customer experiences, ensuring that the right content, offers, or communications are delivered to the user at the most opportune moment in their journey.

The ODE DecisionEvents Dataset tracks decision events and proposition outcomes in Adobe Journey Optimizer. It focuses on offer propositions made to users, tracking how decisions are made within the system and how users interact with those propositions. This dataset is used to understand the performance of decisions, offers, and how users respond to them. It is crucial for reporting and analysis around decision-making processes, offer performance, and user engagement with propositions.

Key Use Cases for the ODE DecisionEvents Dataset:

1. Offer Performance Tracking: Track how users engage with offers, including clicks, views, and conversions, to optimize offer strategies.

2. Decision-Making Analysis: Analyze how decisions are made based on rules, algorithms, or strategies, and measure the performance of decision options.

3. Customer Experience Personalization: Monitor how personalized offers and experiences are delivered based on user profiles and journey interactions.

4. Optimization of Decision Strategies: Improve decision-making processes by analyzing the performance of proposition strategies, algorithms, and fallback options.

5. Experience Outcome Measurement: Capture outcomes based on decision events, including success, failure, or other actions that reflect user engagement with propositions.

Relationship Between ODE and AJO Entity Dataset:

1. Linking via Journey Structure:

   • The AJO Entity Dataset tracks the entire structure of a journey, including journey steps, messages, and decision points.

   • Decision points in the journey are where the ODE Dataset comes into play. When a decision needs to be made, such as which offer to present to the user, the decision event is logged in the ODE Dataset.

   • The AJO Entity Dataset would include references to these decision events, ensuring that every decision made in the journey is tracked.

2. Offer Propositions and Decision Tracking:

   • Offer decisions made during a journey are recorded in the ODE Dataset, which tracks proposition offers and their outcomes (e.g., which offer was selected and how the user interacted with it).

   • These decisions are triggered as part of a journey step in the AJO Entity Dataset, where a decision point is encountered. The AJO Entity Dataset logs the context around why a decision was needed, such as user segment data or behavior during the journey.

3. Common Identifiers:

   • Both datasets share common identifiers such as Journey IDs, Message IDs, and Decision IDs that link the decision events in the ODE Dataset back to the specific journey steps in the AJO Entity Dataset.

   • For example, a Journey ID in the AJO Entity Dataset would link to a decision event in the ODE Dataset, showing how a decision was made within that journey and what offer was presented to the user.

4. Decision Outcomes and Journey Actions:

   • Once an offer decision is made (logged in the ODE Dataset), the outcome of that decision (e.g., user accepts or ignores the offer) is tracked as part of the user’s journey.

   • The AJO Entity Dataset would log the overall journey progress, while the ODE Dataset provides the specific outcome of the offer decision and whether the user engaged with it. This provides a full picture of how decisions affect the user’s journey.

5. Optimization and Personalization:

   • The ODE Dataset feeds back into the AJO Entity Dataset by providing insights into which offers work best for certain segments of users. This data can be used to optimize future decisions within the journey.

   • For example, if the ODE Dataset shows that certain offers are leading to high engagement rates for a specific segment, the AJO Entity Dataset can trigger those offers more frequently during similar journey steps.

Why Track All Proposition Offers and the Algorithm Used?

1. Optimize Offer Strategy and Personalization

• Offer propositions are often personalized based on a user’s profile, behavior, or journey step. Tracking all proposition offers allows marketers to analyze which offers resonate most with specific segments of their audience.

• Algorithms play a central role in deciding which offer or experience is presented to the user. By tracking the algorithms used, you can evaluate how effective each decision-making method is in delivering the right offers.

• Example: If you are running a personalized journey with different product recommendations, tracking which offers are being presented (and the underlying decision logic) lets you fine-tune those recommendations based on engagement outcomes.

2. Measure Offer Performance and User Engagement

• Tracking all offer propositions allows you to measure how well different offers perform in terms of engagement. For example, tracking metrics like click-through rates (CTRs), conversions, or acceptances of offers provides insights into which offers are driving desired behaviors.

• By monitoring the proposition outcomes, you gain insight into how different types of users respond to various offers. This helps in identifying trends, such as which offers lead to higher engagement with a certain demographic or segment.

• Example: Suppose you are running a campaign with multiple offers (e.g., discount codes, product recommendations). Tracking which offer users engage with (e.g., accepting a discount code vs. ignoring a recommendation) helps you adjust the future decision-making process to favor more successful offers.

3. Test and Improve Decision-Making Algorithms

• Algorithms determine which offers are presented to a user. Different algorithms may prioritize different factors (e.g., recency of interaction, likelihood of conversion). Tracking which algorithm was used for each decision allows you to evaluate the effectiveness of various decision-making strategies.

• Why it matters: Not all algorithms will work equally well for all users. For example, an algorithm based on past behavior might work better for returning customers, while a rules-based algorithm might perform better for new users. By tracking the algorithm's performance, you can refine the decision-making process and tailor it to specific contexts.

• Example: You may use one algorithm to optimize for maximizing engagement and another for driving conversions. By tracking how each algorithm performs under different conditions, you can choose the best one for each scenario.

4. Understand Fallbacks and Avoid Missed Opportunities

• Sometimes, none of the primary offers may meet the decision criteria, so a fallback offer is presented to avoid presenting no offer at all. Tracking the fallback mechanism ensures you understand when your primary offers are insufficient and that you don’t miss opportunities to engage users.

• Example: If you find that fallback offers are being used frequently, it may indicate that your decision-making process needs optimization. Maybe your primary offers aren’t relevant enough, or the targeting rules are too restrictive. By tracking the use of fallback options, you can adjust your strategy to improve primary offer performance.

5. Support A/B Testing and Iteration

• Tracking all offer propositions and the algorithm used allows for A/B testing of different decision strategies. By analyzing which offers (and which decision algorithms) yield the best engagement or conversion results, you can iteratively refine and improve the customer journey.

• Example: Suppose you're testing two different algorithms—one that prioritizes discounts and another that prioritizes recommendations. By tracking the propositions and outcomes, you can determine which approach leads to better results for specific segments, then optimize your future campaigns accordingly.

6. Improve Customer Experience

• By tracking proposition outcomes, you ensure that users receive the most relevant and timely offers. This helps maintain a consistent and personalized customer experience, leading to higher satisfaction and loyalty.

• Why it matters: Presenting irrelevant offers or poorly timed propositions can degrade the customer experience. Tracking helps prevent this by ensuring you present the best possible offer or take corrective actions when engagement is low.

• Example: If a user consistently ignores product recommendations but engages with discount offers, tracking the decision events allows you to tailor future offers to align with their preferences, improving the overall experience.

Sample Queries

Retrieve User Information Along with Proposition Offers

SELECT to_json(p.identityMap), 
       to_json(p.proposition) 
FROM (
  SELECT identityMap,
         EXPLODE(_experience.decisioning.propositionDetails) AS proposition 
  FROM   ode_decisionevents_example_decisioning
) p

This query extracts user identity information and proposition details from the ode_decisionevents_example_decisioning dataset. It works by first selecting the identityMap (which contains user identity data) and exploding the propositionDetails array (which holds details of propositions made to users) so that each proposition is returned as a separate row. The outer query then converts both the user identity and the proposition details into JSON format, making them easier to work with for further analysis or integration into other systems. This approach is typically used to track the specific offers or decisions made for each user during their journey.

Extracting Decision Event Details by Year and Month

SELECT  s.timestamp,
        STRING(YEAR(s.timestamp)) AS year,
        STRING(MONTH(s.timestamp)) AS month,
        STRING(YEAR(s.timestamp) * 100 + MONTH(s.timestamp)) AS yearmonth,
        s.propositionId,
        s.eventType,
        s.customerId,
        s.activityName,
        s.activityId,
        s.selection.name AS offerName,
        s.selection.id AS offerId,
        s.placementName,
        s.placementId
FROM (
        SELECT  p.timestamp,
                p.propositionId,
                p.eventType,
                p.identityMap.customerid[0].id AS customerId, 
                p.proposition.activity.name AS activityName, 
                p.proposition.activity.id AS activityId,
                EXPLODE(p.proposition.selections) AS selection,
                p.proposition.placement.name AS placementName,
                p.proposition.placement.id AS placementId
        FROM (
            SELECT timestamp,
                   _id AS propositionId,
                   eventType,
                   identityMap,
                   EXPLODE(_experience.decisioning.propositionDetails) proposition 
            FROM   ode_decisionevents_example_decisioning
        ) p
) s

This query extracts detailed information from the ode_decisionevents_example_decisioning dataset, focusing on propositions (offers) presented to users. It retrieves fields such as the event timestamp, propositionId, eventType, customerId, activityName, activityId, offer name and ID, and placement details. Additionally, it formats the timestamp to generate year, month, and a concatenated yearmonth field for temporal analysis. The query uses the explode function to break down the array of selections (offers) into individual rows, ensuring that each offer is captured separately. This structure allows for a granular view of the decision events, tracking when specific offers were made and linking them to the customer, activity, and placement involved.

Activity Count by Decision Type

This chart shows the count of activities grouped by decision types.

SELECT 
    proposition.activity.name AS activityName, 
    COUNT(*) AS numOfActivities
FROM 
    ode_decisionevents_example_decisioning
GROUP BY 
    proposition.activity.name
ORDER BY 
    numOfActivities DESC;

Offers Per Placement

This chart shows the number of offers per placement.

SELECT 
    proposition.placement.name AS placementName, 
    COUNT(*) AS numOfOffers
FROM 
    ode_decisionevents_example_decisioning
GROUP BY 
    proposition.placement.name
ORDER BY 
    numOfOffers DESC;

Offers Served Per Month

This chart tracks the number of offers served each month.

SELECT 
    STRING(YEAR(timestamp)) AS year, 
    STRING(MONTH(timestamp)) AS month, 
    COUNT(*) AS numOfOffers
FROM 
    ode_decisionevents_example_decisioning
GROUP BY 
    STRING(YEAR(timestamp)), 
    STRING(MONTH(timestamp))
ORDER BY 
    year, month;

Unique Customers With an Offer Proposition Per Month

This chart shows the number of unique customers who received an offer each month.

SELECT 
    STRING(YEAR(timestamp)) AS year, 
    STRING(MONTH(timestamp)) AS month, 
    COUNT(DISTINCT identityMap.customerid[0].id) AS numOfUniqueCustomers
FROM 
    ode_decisionevents_example_decisioning
GROUP BY 
    STRING(YEAR(timestamp)), 
    STRING(MONTH(timestamp))
ORDER BY 
    year, month;

These queries assume that the dataset follows the structure shown in the previous example. You can adjust column names or logic based on your specific schema or dataset requirements.